Recent years have seeninnovations in foundationconstruction and also innovationsin techniques for solvingfoundation problems.However new procurement methods,especially Design & Build, havefocused attention almost exclusivelyon cost and programme. Lack ofchecks on design and constructionhave led to declining technicalstandards and an increasing numberof foundation failures.This paper reviews currentpractice and highlights some ofthe issues.

‘Because the success or failure of building lies buried deep with the foundations, we ignore them.But if we do, then like a child that grows up and turns round to disappoint us with their behaviour so willthe foundations when we build without forethought and diligence to the ground and materials we use.’

- Karl TerzaghiFoundationsFoundations are simple in principle. There are two main types:* a spread footing connects to the structure, takes the load and spreads it out to reduce the pressure towhat the soil can take.* a piled foundation connects to structure and transfers the load to piles with solid shafts which transmitload down to a strong stratum at depth

Other alternatives are:* vibrocompaction, where stone columns are formed to consolidate loose soil; and* grouting, where grout is injected into the ground under pressure to fill voids.All of these techniques are well established and have been around for many years.

UnderpinningThis involves deepening an existing foundation or parts of foundation down to firmer ground to create evensupport for the structure. There are two main techniques:* mass concrete, or* minipiles supporting the structures via needles or a raft.

Modern developments

Continuous flight auger piles work like a corkscrew: first, the auger is screwed into the ground. and thenslowly pulled up out of the ground, removing a plug of soil with it. The auger has a hollow stem andconcrete or grout is pumped down this as it is withdrawn to fill the hole and form the pile.

There are thenvarious types ofmicropiles:* bored micropiles(e.g. Odex,Ischebeck);* top-driven andb o t t o m - d r i v e nmicropiles.

(Illustrations from MJ Tomlinson Pile Designand Construction Practice, 4th Edition, Spon,London, 1994)

F o u n d a t i o ndesign bibles:* BS8004* Terzaghi &Peck: SoilMechanics inEngineeringPractice*Tomlinson:FoundationDesign &Construction* Tomlinson:Pile Design &Construction

Spread footing design:estimate strength of ground,estimate safe bearing capacitycalculate required foundation size

Pile design:* calculate shaft friction* calculate end bearing capacity* apply appropriate safety factors to determineSafe Working Load (SWL)

The two most widely accepted pile driving formulae are the Hiley formula for top-driven piles and the BSPformula for bottom-driven steel tube piles. In the Hiley formula it is important to quantify all the coefficientsand allowances as accurately as possible, as specified in CP4 (copy in metric units available atwww.anbeal.co.uk). The results from these formulae can be reasonably accurate for short piles in sand orgravel but are less accurate for long piles and accuracy is poor in silt and clay. Therefore they cannot beconsidered a substitute for geotechnical design. However they are useful for calculating the size of hammerrequired and also monitoring pile driving on site.

Testing* According to CIRIA report R144,

standard pile integrity tests onlyreliably ‘see’ a pile length ofabout 20-30 diameters fromground level, so caution is neededwith long small diameter piles.

* A static load test (carried out inaccordance with the ICE pilingspecification) is expensive andinconvenient to the contractor butit is the only sure way to checkpile load capacity.

* Indirect tests using CAPWAP andSIMBAT are useful and can bereasonably accurate for shortpiles but are less accurate for longpiles and have limited accuracyin clay, as they do not applysustained load.

Work below ground involves many unknowns. The engineer needsadequate ground information to understand how the foundation isworking and what is below it. For piles, boreholes should typicallypenetrate at least 10m deeper than the expected pile depth anddeeper still in mining areas.Driving piles ‘blind’ is a formula for disaster - the contractor doesnot know what is generating the driving resistance, or what isbelow the toe of the pile. In this situation, the claimed load capacityof the pile is pure guesswork.* Bored piles can suffer loss of section in

soft or wet ground.* The capacity of driven piles in clay can

change as porewater pressures changeafter they are driven, so redriving maybe necessary.

* The capacity of piles in groups may bereduced.

* Other possible sources of trouble couldbe a deep peat layers, or a layer of gravelover clay.

* Where compacted made ground has beenplaced over soft soil or fill, this can settleover time, so that instead of contributingto pile load resistance it applies adowndrag force to the pile shaft.

* The early years of CFA piles sawmany problems, now mostly solvedby precise instrumentation andmonitoring of boring andconcreting. However the quality ofthe pile base can still be unreliable.In wet and soft soil, ‘flighting’ canoccur: the auger pumps soil out ofthe hole instead of drilling into it(like a corkscrew in a bad cork).

* Segmental Flight Augers are nowused for smaller diameter and workin limited headroom. Manyproblems can occur with these, asmany rigs are not fully instrumentedand the boring and concretingprocesses are not continuous.

* BS8004 says that piles should be made of concrete. However smalldiameter piles are often formed from 1:1 or 1:2 grout. Shrinkage andalkali silica reaction (ASR) are both potentially serious problems withthese mixes and their structural performance is uncertain. An enquiryto the Building Research Establishment revealed that they wereunaware of any research into these issues. They were also not awarethat it had become common practice to form structural columns (piles)out of sand/cement grout. Beware!

* Underpinning must be taken down to a firmstable base and it must be connected to structure- it is no use if it is not properly pinned up. Ifthe concrete is ‘poured under a head’ instead ofbeing pinned up, unless care is taken theunderpinning may not support the structure atall.

* If clay shrinkage is the problem, piled underpinningis usually unsuitable, as it creates a two-levelfoundation. Unless precautions are taken, the buildingmay be lifted off the piles when the clay swells inwinter.

Grundomat minipiles - steeltubes driven by a vibratinghammer, typically 150mmdiameter. On this size of pile,most of the surface area is theshaft, so most of the loadcapacity comes from frictionand only a small amount fromend bearing. When these pilesbear on weak rock they canhave very low capacity.

Unorthodox systems

One-sided piling: the attraction of the idea is understandable but what happens if the external walls of ahouse are piled and internal walls are not? Differential movement could be a serious problem. Also thereare technical problems:* piles driven to refusal may have very low tension capacity unless they penetrate a long way into a firm

stratum;* a system based on raking piles will generate lateral forces - and unless the ground is good enough to

resist this, movement may occur; the sharing of forces between the piles will depend on where this forceacts.

* The load capacity of one-sided knuckle piles is limited by the moment capacity of the pile and its headconnection.

* One sided screw piles: pile capacity will depend on soil conditions and also on the moment capacity ofthe pile and its head detail. It will also depend on the strength of the connection to the existingfoundations. The load capacity when all of these are considered is likely to be very low. Where theproperty has cavity walls how is the inner leaf of a cavity wall supported?

* Pile jacking can be auseful technique.However note that thepiles should be testedto 1.5xSWL (SafeWorking Load) toprove their adequacy.If the piles are in claythe load must beapplied for a sustainedperiod to bemeaningful - whichprobably makes thetechnique impracticalin this situation.

(Illu

stra

tions

from

MJ T

omlin

son

Pile

Des

ign

and

Con

stru

ctio

n Pr

actic

e, 4

th E

ditio

n, S

pon,

Lond

on, 1

994)

* Although installation of the‘Hoopsafe’ system looks likeunderpinning, with holes beingexcavated around the property andfilled with steel and concrete, it is infact very different: the concrete goesalongside the existing walls andfoundations, rather than below them,and it is prestressed on completion.When considering possible use of iton a project, the engineer needs toconsider how it will work with theexisting structure. Is the idea toprestress the existing brick walls orto prestress the new concrete beams?What happens to the internal walls?How strong is connection betweenthe Hoopsafe beams and the housemasonry walls? How is the inner leaf of a cavity wall inner leaf supported?The suitability of the system for a particular project will depend on the answers to these questions.

* Expanding grout injected into the ground is promoted as an alternative to underpinning. One problemwith grout is that it goes where you want it to, also where you don’t want it to (e.g. drains, cellars ofadjacent properties). The representative for a firm selling one system at a seminar told us in allseriousness that it had a special molecular structure which expands vertically but not horizontally!

* On the project illustrated above, insurers insisted on expanding grout instead of underpinning for anextension which was settling relative to the main property. Movement was monitored before andafter - and it moved faster after grouting than it had before. Under their 10 year guarantee thecontractor offered to pump in more grout. We decided to underpin it properly instead. As can beseen, the grout had spread horizontally as well as vertically (surprise?). The soil was fill comprisingsoft clay and sand and the grout had run into the sand layers but not the clay. Compressing soft clayonly makes it worse, so it was no surprise that the expanding grout failed to solve the problem.

* Precast concrete pilesare usually connectedtogether by costlyfabricated joints but onemanufacturer has asystem held together bylocating dowels andglue. The trouble is thateach hammer blowsends a shock wavethrough the pile whichcan send the endsegment flying off (likethe balls in ‘Newton’scradle’), unless theground is very firm.‘The house that sank’was a new house onthese piles that settledby 1 metre.

* On a recent project a major national contractor used SFApiles. He did the integrity testing himself but several pilesfailed. The specified SWL was 310kN and a static load testto 1.5SWL. The contractor argued noisily against static loadtesting, saying that it cost too much and ‘nobody does thisany more’. The first test passed 1.0 SWL successfully butat 1.25SWL it settled 26mm - failure. A second pile settled22mm at 1.5SWL and kept moving, so it failed too. Therewere no more arguments about static load tests. Integritytests by an independent testing contractor found twice asmany defective piles as the piling contractor’s own testing.The whole job had to be re-piled.

As noted earlier, the BSP and Hiley formulae have limited accuracy is limited but are fairly respectable.* A typical formula for

Grundomat piles predictsinfinite load capacity at refusal!I have seen failures ofGrundomat piles driven to rock.Geotechnical calculations showend-bearing capacity is smalland pile loads should be modest.

* The Engineering News formuladates from 1882 and uses asafety factor of 6. In the 1960sTerzaghi and Peck said it shouldno longer be used.

* SCI publication P156 includesthe EN formula, saying it is usedin the USA. However it fails tomention that it needs a safetyfactor of 6 and is hopelesslyinaccurate - and it gets theformula wrong.

Some anecdotes about real problems(i) A national piling contractor who believed driven piles

are ‘self testing’ calculated the set for each piledepending on its load and drove long piles in clay withlayers of peat on this basis. When some failed to achievetheir designed set, he calculated a revised capacity basedon the actual set and added more piles. What he did notrealise was that (a) the Hiley formula is not accurate forlong piles in clay and (b) the failure to achieve a set wassign of a soft layer (peat) at depth. The piles failed.

(ii) On a job in Cheshire a contractor proposed 200mmbottom driven piles with SWL = 200kN. Based ongeotechnical calculations, the estimated pile length was 13-14m. The piling contractor said this wasvery conservative and the piles would achieve refusal well before this. I asked him what weight ofhammer he used - he said 250kg, with a 1.7m drop. With this the BSP formula predicted a refusal loadRu = 154kN, so no wonder the piles couldn’t be driven to 13-14m - they were never going to achievethe specified SWL of 200kN. Because the piling contractor did not know about the BSP piling formula,he had bought too light a hammer.

(iii) On another job a contractor proposed 220 diameter driven piles for 320kN SWL. Instead of using astandard formula, their engineer relied on a formula his old boss had devised which predicted over1500kN ultimate capacity. The standard BSP formula predicted less than half this. The contractorpromised that the piles would be OK. He was confident they would easily carry the load and arguedthat a load test was expensive and unnecessary. However on site the piles struggled to achieve the designset. The load test failed at about 175kN i.e. 50% of the specified SWL. There were no more argumentsabout load testing - the job had to be re-piled.

(iv) On an underpinning job in Scotland, with fill overlying clay and rock, we told the contractors that boredpiles were probably appropriate but one contractor proposed:

(i) Grundomat piles bearing on rock, which were rejected, then (ii) pile jacking to 1.0SWL, which was rejected, then (iii) pile jacking to 1.5SWL but no sustained proof load, which was rejected because of the clay, then (iv) driven piles with a predicted 300kN SWL for 220mm piles driven with a 0.5t hammer. The contractor

used the SCI P156 Bearing Piles formula (see earlier), which predicted Ru = 785kN and SWL = 314kNwith a safety factor of 2.5. However the standard BSP formula predicted Ru = 161kN and SWL 64kN,so this proposal was rejected.

The contractor then gave up suggesting different ideas and accepted that bored piles would have to beused.

I could go on ....It is clear that design to BS8004 is the exception rather than rule. Many piling contractors are simply guessingload capacity, driving piles to near refusal and hoping for the best. ‘Who Dares Wins’ - the piling contractorwho uses fewest piles bids the lowest cost and wins the contract.It is important to remember that, when a consulting engineer is appointed to design foundations for a building,it is not enough to simply state that the piles should be designed by a specialist contractor to carry the statedloads. In a recent case in Northern Ireland, the judge (rightly) ruled that in this situation the consultingengineer still had a duty to his client for the adequacy of the foundations. Although the engineer could notbe expected to carry out the full detailed design of the piles, he had a duty to carry out approximate checksto satisfy himself that the design was ‘in the right ballpark’.When foundations fail, the costs to the project are out of all proportion to the money saved by skimping onthe design and construction. ‘Leaving the design to the contractor’ can be a formula for disaster. For theirown good and also the good of their clients, engineers need to assert their authority more on projects, carryout simple checks on designs submitted by piling contractors - and reject those which are not good enough.For details of pile driving formulae and other engineering issues, see my website www.anbeal.co.uk

Alasdair Beal March 2012